Research On Control Of Patient-dominated Assisted Manipulator Based On Hand Motion Prediction

Hand is one of the most important organs of human being.It has the characteristics of multi freedom and flexible movement.On the one hand,it is difficult to obtain the complete and continuous fine movement of the hand.On the other hand,it is difficult to recover the movement of the hand after the hand is injured or the biomechanical function and neurophysiological characteristics of the hand muscles are changed due to stroke.In recent years,great progress has been made in the research of humanoid manipulators.A batch of humanoid manipulators with full drive and under drive have emerged,which are used in many fields,such as industrial manufacturing,humanoid robot design and prosthetic hand,with the characteristics of small weight,simple structure and powerful function.For the patients with hand movement disorders,if the humanoid manipulator can help them to complete some daily tasks,it will bring great convenience to life,and the underactuated manipulator with lower cost and more convenient use is more suitable for the patients with hand movement disorders.The existing research often focuses on the design and fine control of underactuated manipulator,neglecting the auxiliary control research with the patient as the main object.Based on the prediction of hand movements,this paper studies the optimal underactuated manipulator control based on the cooperative base of hand gestures.At the same time,it takes the patient as the main body,predicts the movement intention of the patient,and realizes the patient dominated auxiliary control of the controlled manipulator.The main research work of this paper is as follows:(1)Hand modeling and kinematics analysis.According to the anatomic structure,the complete joint model of human hand is built,and the joint model of human hand is simplified.The forward and reverse kinematics models of each finger are analyzed,and the forward kinematics models based on D-H parameters are established for different fingers,based on which the complete joint hinge model of the hand is established.Based on Jacobian matrix,the inverse kinematics model is established,and the inverse kinematics simulation of fingers is carried out,which is the basis of studying hand motion and solving the position and pose of underactuated manipulator as end effector.(2)The research of hand gesture cooperative base extraction and the simulation of underactuated manipulator grasping.This paper analyzes the movement data of hand grasping,establishes the coupling relationship between the joints of fingers,and proposes a method to extract the cooperative base of hand gestures,decomposes and simplifies the grasping movement of hand,reduces the dimension of grasping movement data,and corresponds to the control input of underactuated manipulator,so as to improve the real-time and accuracy of the control of underactuated manipulator.On this basis,the grasping model of underactuated manipulator is studied,and the grasping control of underactuated manipulator based on gesture coordination is planned and simulated.(3)Prediction of hand motion and auxiliary control of underactuated manipulator.According to the existing classification method of hand grabbing action,according to the needs of patients,the commonly used grabbing actions are selected to be subdivided,and an algorithm based on convolutional neural network(CNN)is proposed to predict the motion intention of patients’ hands.The motion capture system is used to collect the movements of hands before the grabbing action is completed,predict the grabbing task that patients will complete,and predict the movement intention of patients The measurement results control the underactuated manipulator to complete the grasping task,and realize the auxiliary control of the underactuated manipulator led by the patient.